In recent years, machine learning has been increasingly utilized in speech and language processing. For example, deep learning techniques may be used in phone recognition, speech feature coding, and large vocabulary speech recognition. For example, a goal of spoken language understanding (SLU) in human/machine spoken dialog systems may include automatic identification of the domain and intent of a user as expressed in natural language, and extraction of associated arguments/concepts or slots to achieve a goal in SLU for human/machine dialogs.